Machine Learning Analysis of Gene Expression Reveals TP53 Mutant-like AML With Wild Type TP53 And Poor Prognosis

Abstract Acute myeloid leukemia (AML) with TP53 mutations (TP53Mut) has poor clinical outcomes. We investigated whether this AML subtype harbors a distinct gene expression profile (GEP) and whether this GEP is prognostic in TP53 wild type (TP53WT) AML. We applied a supervised machine-learning approach to detect the TP53Mut GEP. We divided the samples in the Beat-AML dataset into training and testing datasets. The TCGA LAML dataset was used as a validation dataset. We trained a ridge regression machine learning model to classify TP53Mut and TP53WT cases. This model was highly accurate in distinguishing TP53Mut versus TP53WT cases in both the test and validation datasets. Additionally, we identified TP53WT samples with high ridge regression scores. These high-scoring TP53WT samples also have poor overall survival, suggesting they share clinical and GEP features with TP53Mut AML. We defined these TP53WT samples as TP53Mut-like. Using drug sensitivity data in the Beat AML dataset, we found TP53Mut-like AMLs have distinct drug sensitivity patterns that phenocopy TP53Mut cases. Finally, we identified a 25-gene signature that can identify TP53Mut-like AML. This signature could be used clinically to identify this novel subset of poor-prognosis AML.